Previous factor analyses of data from the Dunedin (New Health and Development Study have identified three dimensions of temperament shown to predict later social development. The purpose of this proposed study is to develop a laboratory-based procedure for assessing these dimensions of temperament. Subjects will be 50-60 male and 50-60 female children between 3 and 5 years of age, who will participate in two testing sessions, scheduled approximately 2 weeks apart. The children will be videotaped of the children's behavior during the sessions will then be rated on a series of items which comprise scales labelled "Lack of Control", "Approach/Sociability", and "Reaction to Novelty/Wariness"; additionally, items comprising a fourth scale, "Activity Level" will be included. Parent ratings of the child's temperament, behavior problems, emotionality, and family environment will be collected. The four laboratory-based temperament scales will be evaluated in terms of internal consistency, inter-rater reliability, correlations with parent reports of child characteristics, and correlations with cognitive characteristics assessed during the testing sessions.